December 7th

Anthropic acquired JavaScript runtime Bun after announcing that Claude Code reached $1 billion in run-rate revenue just six months after launch—underscoring massive enterprise demand for agentic coding solutions and positioning Anthropic to vertically integrate its development stack.

Harvey raised $160 million at an $8 billion valuation, more than doubling since February. Coupled with $150 million in annual recurring revenue, this signals that vertical-specific AI applications are successfully moving beyond hype to substantial commercial adoption.

Microsoft lowered sales growth targets for its AI agent products by 25% to 50% in some units after salespeople missed quotas, suggesting that despite the "agentic era" narrative, enterprise customers remain cautious about deploying autonomous agents.

Replit secured a multi-year partnership with Google Cloud after hitting $150 million in annualized revenue, giving Google a strategic foothold against Microsoft's GitHub Copilot by targeting the "vibe-coding" market of non-traditional developers.

Black Forest Labs raised $300 million at a $3.25 billion valuation to expand its Flux image generation models, with strategic investors Canva and Figma indicating consolidation of generative visual AI into professional design workflows.

Gong released a study showing sales teams using AI generate 77% more revenue per rep than non-users, with 70% of revenue leaders now trusting AI for decisions—quantifying the shift from experimental AI to critical revenue infrastructure.

AWS launched "Frontier Agents" (Kiro, Security, and DevOps) capable of working autonomously for days to solve complex coding and operational problems, signaling a shift toward AI replacing entire workflows rather than just augmenting tasks.

Nature reported that scientists using AI tools publish more papers and become team leaders four years earlier than their peers, creating a significant career velocity gap that pressures researchers to adopt AI to remain competitive.

Amazon's AI chatbot Rufus drove an 805% year-over-year increase in AI-driven traffic to retail sites on Black Friday, with sessions involving the chatbot 100% more likely to result in a purchase—validating AI as a high-conversion sales channel.

The New York Times filed a lawsuit against Perplexity, alleging the startup's "answer engine" creates verbatim copies of copyrighted work to siphon revenue—a critical test for whether AI search tools can legally bypass publisher paywalls under fair use.

Anthropic released a 153-page system card for Claude Opus 4.5, revealing a 0% attack success rate after 200 attempts in computer use scenarios, highlighting a divergence in safety transparency compared to competitors' shorter documentation and higher vulnerability rates.

Meta faces an EU antitrust investigation regarding its policy of restricting third-party AI access on WhatsApp, signaling that regulators are moving quickly to prevent platform gatekeepers from monopolizing the AI application layer.

The US Department of Justice charged two contractors for attempting to wipe government databases, noting they used an AI tool to ask how to delete logs. The AI's failure to provide effective cover-up instructions ironically led to their detection, illustrating current limitations in AI-facilitated cybercrime.

OpenAGI emerged from stealth with Lux, achieving 83.6% success on the Online-Mind2Web benchmark—significantly outperforming OpenAI's Operator (61.3%) and Anthropic's Computer Use (56.3%), suggesting smaller specialized architectures may be superior to generalist LLMs for autonomous computer control.

NVIDIA and the University of Hong Kong released Orchestrator, an 8-billion-parameter model that manages tools more accurately than larger frontier models by using reinforcement learning to delegate tasks, offering a more efficient path for enterprise reasoning systems.

Runway launched Gen 4.5, claiming the No. 1 spot on the Video Arena leaderboard ahead of Google's Veo 3 and OpenAI's Sora 2, demonstrating that specialized startups can still outpace tech giants in specific modalities.

Google's Gemini 3 Pro achieved a 69% trust score in blinded testing compared to Gemini 2.5's 16%, suggesting Google is effectively closing the gap in user alignment and safety perception against competitors.

OpenAI introduced "Confessions," a training method that compels models to self-report misbehavior in a structured post-response report—aiming to mitigate "reward hacking" where models deceive users to maximize positive feedback during reinforcement learning.

Micron announced it will cease selling consumer memory under its Crucial brand to reallocate capacity to AI data center products, highlighting the acute supply chain pressure created by modern AI accelerators' memory requirements.

AWS unveiled Trainium3 UltraServer, which is 4x faster and 40% more energy efficient than its predecessor, with plans for Trainium4 compatibility with NVIDIA's NVLink—aggressively positioning its custom silicon as a viable, lower-cost alternative to the H100 ecosystem.

NVIDIA invested $2 billion in Synopsys to accelerate transitioning chip design software from CPUs to GPUs, tightening its grip on the semiconductor stack and ensuring its hardware remains essential for designing next-generation chips.

Amazon data centers in Morrow County, Oregon, have been linked to nitrate contamination in drinking water, correlating with increased cancer and miscarriage rates—underscoring the growing environmental and public health liabilities of AI infrastructure's physical footprint.